NeilBrown [Wed, 3 Nov 2021 02:08:45 +0000 (13:08 +1100)]
oceani: add lots of UNTESTED markers.
These serve to remind me what needs testing (or marking NOTEST), and
will make it easier to notice when new untested code is added.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 3 Nov 2021 01:52:08 +0000 (12:52 +1100)]
oceani: var storage needs to be free last
As type may store initializers in storage, but these might not get into
the var list in an error, we need to free that storage last.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 3 Nov 2021 01:29:34 +0000 (12:29 +1100)]
oceani-tests: continue to improve test coverage
Add a couple of error tests, and mark some code as NOTEST
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 3 Nov 2021 01:05:51 +0000 (12:05 +1100)]
oceani-tests: add some NOTEST markings.
These lines don't need to be tested.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 3 Nov 2021 01:01:17 +0000 (12:01 +1100)]
mark-untested: improve filter
Don't add the mark for any section lines.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 3 Nov 2021 00:57:01 +0000 (11:57 +1100)]
oceani: add test for duplicate main functions
Add a test so the check for duplicate main functions gets exercised.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 3 Nov 2021 00:46:54 +0000 (11:46 +1100)]
oceani: drop "program" in favour of "func"
Being to introduce functions. Currently only the 'main' function which
receives an array of strings.
This array has a variable size.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 30 Oct 2021 04:49:08 +0000 (15:49 +1100)]
oceani-tests: add test for declaring a CondScope variable
If a variable was declared in all branches of a structures command, it
may or may not be declared as something else afterwards.
We need to test both options.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 30 Oct 2021 04:25:40 +0000 (15:25 +1100)]
oceani-tests: add test for unknown field name.
Test that we get error if unknown field is requested.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 30 Oct 2021 04:19:39 +0000 (15:19 +1100)]
oceani: improve test coverage
Add a tool 'mark-untest' which adds comments to untested lines.
Then resolve some of them, either adding more tests, or marking
lines as NOTEST
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 30 Oct 2021 04:00:03 +0000 (15:00 +1100)]
oceani-tests: make it possible to skip valgrind tests.
valgrind tests are slow and not always wanted, particilarly when
examining coverage. So make it easy to skip them.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 30 Oct 2021 03:23:46 +0000 (14:23 +1100)]
ocean: change program to receive argc and argv
Rather than a list of strings, a "program" now receives and arg count,
and an array of strings with that size.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 17 Oct 2021 10:03:01 +0000 (21:03 +1100)]
oceani: move variable values to a stack frame.
We have two frames - one for global values (currently always constant)
and one for local variables.
When we get functions, the local variable frame will be managed with a
stack of frames.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 17 Oct 2021 09:29:16 +0000 (20:29 +1100)]
oceani: fix some incorrect section headers.
section header that just introduce code are 6 deep, not 2.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 17 Oct 2021 02:45:18 +0000 (13:45 +1100)]
oceani: clean up interp_prog()
In particular, use dup_value() to save the command arguments.
But also re-arrange code to make steps a bit cleaner.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 17 Oct 2021 02:35:58 +0000 (13:35 +1100)]
oceani: add parse_context arg to all interp functions, and a few others.
When I switch variables to use a stack frame, I'll need the
parse_context available more broadly (as it will hold the stack).
So add it to a selection of functions now.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 16 Oct 2021 05:58:42 +0000 (16:58 +1100)]
oceani: differentiate static-sized arrays from others.
Some arrays will always have the same size - a static size.
Others might have a different size each time their scope is entered, if
the size is calculates from a variable.
The latter need to be reallocated whenever scope is entered, the former
do not.
This will matter when we create call frames to be able to handle
recursion.
So detect and handle the difference now.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 16 Oct 2021 05:27:41 +0000 (16:27 +1100)]
oceani: don't allocate init value for non-initialized fields.
Struct fields that aren't explicitly initialised must be initialized to
a 'null' value. This can happen at interp-time. There is no need to
allocate a null value when parsing.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 16 Oct 2021 05:09:59 +0000 (16:09 +1100)]
oceani: implement struct field initialisation properly.
I wasn't testing default initial values for struct fields, so of course
it didn't work.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Thu, 14 Oct 2021 02:59:34 +0000 (13:59 +1100)]
oceani: move ->prepare_type call (back) into val_alloc()
This ensures it is called for struct fields as well.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Thu, 14 Oct 2021 02:43:02 +0000 (13:43 +1100)]
oceani: handle variable-sized arrays better.
An array with size set by a constant variable(!) might have a different
size each time the declaration is encountered. So we need to
re-evaluate the size each time.
We currently re-evaluate the size only if it is zero.
So for numerical-constant sized arrays, evaluate size during parsing.
For other arrays, re-evaulate each time using a new prepare_type method.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Tue, 12 Oct 2021 10:28:47 +0000 (21:28 +1100)]
oceani: fix a couple of issues
1/ when a variable declared in a loop was re-initialized, we didn't free
the old value before allocating a new one.
2/ When assigning to an out-of-bounds array index, created an rval,
but never freed it.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 9 Oct 2021 21:14:04 +0000 (08:14 +1100)]
oceani - discard anon_typelist
Don't keep anon types on a separate list, use the same list but given
them an empty name.
This allows us to use add_type() for adding all types.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 9 Oct 2021 01:05:51 +0000 (12:05 +1100)]
oceani: drop 'array' field from the 'value' union.
This field was never used - we use the generic 'ptr'.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Thu, 7 Oct 2021 00:13:44 +0000 (11:13 +1100)]
parsergen: require all terms to be declared if any are.
The previous implementation only required all terms used for the first
time after a $TERM declaration to be declared. That isn't really
useful.
So change it to only set the default symbol type of Term after all the
grammar is parsed, and then only if not $TERM declarations were found.
This highlights that various terminals in ocean weren't declared, so
declare them now.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 2 Oct 2021 22:57:40 +0000 (09:57 +1100)]
oceani: drop parse_value method for types.
The args to 'program' are now all strings. The program itself needs to
convert to numbers etc as needed.
So we don't need 'parse_value' any more.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 2 Oct 2021 22:36:50 +0000 (09:36 +1100)]
ocean: introduce prefix op for string->number conversion.
Rather than having magic conversion of command line args to numbers as
needed, introduce '$' as a prefix op to to the conversion.
This is a step towards changing 'program' to be a 'main' function.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 29 Sep 2021 22:21:15 +0000 (08:21 +1000)]
oceani: separate types out from values
Rather than embed the type in ever value, keep it separate.
This allows arrays without duplicating the member-type information.
Also review and clean up the commentary.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Tue, 6 Apr 2021 22:02:20 +0000 (08:02 +1000)]
parsergen: add missing {}
These have ALWAYS been missing, yet the code mostly worked. Weird.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 10 Mar 2021 05:14:20 +0000 (16:14 +1100)]
parsergen: doco updates
Minor revision to descriptive text following all the recent changes.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 10 Mar 2021 04:26:59 +0000 (15:26 +1100)]
parsergen: don't bother with indent_depth
We don't need to store the indent_depth if we just shift the bit-stack
up and down. <<1 pushes a zero, >>1 pops, |=1 changes the zero to one.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 10 Mar 2021 01:37:46 +0000 (12:37 +1100)]
oceani: updates for new approach to parsing indents.
Now that IN is a valid stand-alone token, it makes sense to change the
grammar for ocean.
We don't need the ':' before an indent if there is some other terminal
there. So:
while
statements
do
statements
doesn't require any ':'.
We use the ':' to separate an expression from following statements,
in 'if' and 'while' and 'case'.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 10 Mar 2021 01:13:05 +0000 (12:13 +1100)]
Update indent_test grammer to align with new approach
Indents are now now handled very differently. Update itest
grammar to match.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 10 Mar 2021 00:49:24 +0000 (11:49 +1100)]
parsergen: add support for EOL token
And EOL token is generated when a NEWLINE is found and an EOL can be
shifted. This allows a product to declare that it must finish at the
end of a line, without consuming the NEWLINE.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 10 Mar 2021 00:38:55 +0000 (11:38 +1100)]
parsergen: implement new handling of IN/OUT and NEWLINE
IN/OUT are now expected in the grammar.
In a state where an IN can be shifted, IN symbols are significant to the
grammar. IN symbols appearing anywhere else are ignored (except for how
they affect NEWLINEs).
OUT symbols are ignored precisely when the matching IN was ignored.
NEWLINEs are ignored if the most recent IN was ignored, otherwise they
are significant for the grammar.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 10 Mar 2021 00:12:50 +0000 (11:12 +1100)]
parsergen: split out heart-of-the-parser code
This finishes the break up of the core parser code.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 10 Mar 2021 00:00:23 +0000 (11:00 +1100)]
parsergen: split out the "shift or ignore" section of parsing.
This will make it easier to document.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Tue, 9 Mar 2021 23:56:15 +0000 (10:56 +1100)]
parsergen: split out reduce step of parser
Split out the "reduce" code so it can be more easily documented.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Tue, 9 Mar 2021 23:51:19 +0000 (10:51 +1100)]
parsergen: split out error handling.
I want to split up the core parsing into the different parts so they can
be documented better.
Firstly: split out error handling.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 10:24:14 +0000 (21:24 +1100)]
parsergen: add support for "special" terminals.
We will want a new terminal "EOL", which is like "NEWLINE", but
different. There is currently no room in the numbering for something
like that, so make some room.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 10:26:55 +0000 (21:26 +1100)]
parsergen: remove special case of newline conflicts
newlines are no longer interesting in avoiding conflicts, so remove the
special case handling.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 10:01:17 +0000 (21:01 +1100)]
parsergen: remove tracking of indents and line starts etc.
We will want some of this back, but in a very different form.
So remove it all for now.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 09:41:47 +0000 (20:41 +1100)]
parsergen: remove newline_only
newline_only is not wanted any more. It wasn't being set anyway.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 09:37:34 +0000 (20:37 +1100)]
parsergen: remove starts_line and min_prefix
Neither of these are wanted any more.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 09:31:32 +0000 (20:31 +1100)]
parsergen: remove line_like information.
I'm going to change the 2D nature of the parser over several patches.
First I remove what I don't want, then I add what I do.
During this series, tests won't work!
First: line_like nad relate code is removed.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 26 Feb 2021 06:33:43 +0000 (17:33 +1100)]
parsergen: don't use static buffer for result value.
Add the size of the result value to the per-state information, so it can
be allocated before calling do_reduce(), thus removing the need for a
overly large static buffer.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 24 Feb 2021 08:57:08 +0000 (19:57 +1100)]
parsergen: do not create empty goto arrays
These empty arrays are a waste of space ... assuming they take up
any(?).
Using NULL is more ideomatic.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 10:17:51 +0000 (21:17 +1100)]
parergen: fix bug in deriving itemsets
The code to check if an item had already been added, was wrong.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 10:29:56 +0000 (21:29 +1100)]
parsergen: make "code" arg to "gen_reduce" more obvious.
Call it "pre_reduce" so we know where the code goes.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 10:30:54 +0000 (21:30 +1100)]
parsergen: improve usage message
Now usage message is slightly better.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 22:15:06 +0000 (09:15 +1100)]
oceani: add missing space in usage message.
Wrapping long strings is a bad idea anyway.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 09:20:36 +0000 (20:20 +1100)]
parsergen: fix some problems with choose_sym()
This would sometimes give wrong result, particularly bad when it is
larger than body_size.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 09:18:14 +0000 (20:18 +1100)]
parsergen: move the 'follow' declaration
The 'follow' field is currently used before declared, which is poor
form. Move it.
Also remove a stray '## functions' section that is redundant.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 09:16:22 +0000 (20:16 +1100)]
parsergen: move "Memory allocation" section.
This movement isn't really significant yet, but it will help a bit
later, and I want to get it out of the way.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 09:14:29 +0000 (20:14 +1100)]
parsergen: some missing 'free' calls.
Well... one missing call, and some places were we need to do something a
bit more.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 09:11:04 +0000 (20:11 +1100)]
parsergen: change 'again' to 'check_again'.
Sometimes I use 'again', sometimes 'check_again'.
Change all to use 'check_again'.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 09:07:55 +0000 (20:07 +1100)]
parsergen: assorted updates to descriptive text.
Some of these are fixed for typos or poor grammar. Others update the
text to more closely match the code.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 08:24:11 +0000 (19:24 +1100)]
parsergen.mdc: remove tracking of left-recursive symbols.
left-recursive symbols are no longer interesting, so remove the code and
descriptions.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Wed, 10 Mar 2021 00:58:08 +0000 (11:58 +1100)]
parsergen: guard against tos going negative.
Don't continue the parse loop if tos hits zero.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 08:20:22 +0000 (19:20 +1100)]
parsergen: change how reserved_words are stored
Rather than a simple array with holes, have a dense array mapping number
to name. This will enable a future change which adds names that don't
have numbers assigned.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 08:17:27 +0000 (19:17 +1100)]
parsergen: rename item_index() to item_dot()
The name "dot" fits better with documentation which described the nature
of an "item" when generating LR parsing tables.
Also change the arg to item_num() from 'index' to 'dot'.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 08:04:45 +0000 (19:04 +1100)]
tests: restore coverage testing.
I don't know what changed, but coverage testing stopped working.
Fixed now.
Also add some "|| exit 1" to fail quickly when needed.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 08:00:52 +0000 (19:00 +1100)]
indent-test: enhance the tests a bit.
In particular, test for "else" both indented and unindented.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 5 Mar 2021 07:58:10 +0000 (18:58 +1100)]
indent-test: minor improvements:
1/ add rule to build itest.code
2/ diffs now display correct as "old" and generated as "new"
3/ remove some noise that should never have been there.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 11 Oct 2020 03:49:07 +0000 (14:49 +1100)]
parsergen: add more power to symbol references in generated code
As well as symbol references like "$2", you can now use references
with letters like "$Ss". This will find the shortest symbol in the
production that contains all the given letters in the given order.
There must be a unique shortest symbol.
If that same symbol occurs multiple times, later instances can be given
with a numeric suffix such as "$Ss2".
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 11 Oct 2020 01:18:57 +0000 (12:18 +1100)]
parsergen: update description of $<N
This applies to structures as well as pointers.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 10 Oct 2020 23:34:06 +0000 (10:34 +1100)]
parsergen: allow terminals to be declared.
By default, any non-virtual symbol that does not appear in the head of a
product is assumed to be a Terminal.
For larger grammars, this misses out of an opportunity to detect errors.
So allow a "$TERM" line to list terminals (that do no appear in
precedence lines). If any $TERM line is given, then generate error
if any symbol appears in a production but is not declared, either
as terminal or non-terminal.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 10 Oct 2020 22:50:12 +0000 (09:50 +1100)]
parsergen: avoid infinite loop on error.
If the grammar allows "ERROR" in a recursive location, error handling
can loop for every.
e.g.
foo -> foo bar
foo -> ERROR
Rather than detect and reject such grammars, detect the infinite loop
as it start, and discard an extra token.
i.e. if error handling doesn't discard any tokens from the input
stream, and another error is triggered before anything is shifted, then
we force the next error handling phase to discard at least one token,
or to abort if that token is EOF.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Tue, 6 Oct 2020 07:11:15 +0000 (18:11 +1100)]
parsergen: sort virtual symbols to send of list
virtual symbols are used internally to parsergen
and of no interest to the parser. So they shouldn't
appear in the non_term[] array.
So assign them higher numbers so the nonterminals are dense, and only
include non-virtual non-terminals in the non_term[] array.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Tue, 6 Oct 2020 06:02:22 +0000 (17:02 +1100)]
parsegen: detect left-recursive symbols in non-final position.
A left-recursive symbol that appear other than at the end of a
production causes problem for indent-based parsing, as describe in the
document. So teach parsergen to be able to report them.
Ocean currently has several of these, which I'll need to look into at a
later date.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Tue, 6 Oct 2020 04:44:46 +0000 (15:44 +1100)]
scanner: change the meaning of ignoring comment tokens.
Previously ignoring comment tokens meant they were still parsed, but not
returned. The only way to stop them being parsed was to declare
known marks for the start symbols.
This made is not possible for parsergen to define a language that had
a known mark that would otherwise start a comment.
So change the ignoring of comment tokens to mean they aren't parsed. If
you want to parse comments but not return them, leave the new
"return_comments" field as so. In the unusual case that you want to
return comments set return_comments to 1.
Confirm that this has the desired effect by added in "//" as an
integer-division operator to the sample calculator.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Mon, 5 Oct 2020 23:00:31 +0000 (10:00 +1100)]
indent_test: fix makefile
Maybe 'make' has changed a little to be less forgiving, but 'make itest'
isn't working now. All of LDLIBS are included in the 'cc' line, but
there are no dependencies to make sure they have been built.
The problem is that I'm using LDLIBS for different programs which need
different libs. This isn't such a good idea.
So change indent_test to use itestLDLIBS and itestCFLAGS.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 11 Sep 2020 07:53:49 +0000 (17:53 +1000)]
Rename string.h and number.h headers
Having a "string.h" is potentially confusing as there
is a standard <string.h>
So rename it and for consistency, rename number.h too.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 28 Jun 2019 09:40:11 +0000 (19:40 +1000)]
parsergen: add token location to tracing
When we print the look-ahead token for tracing, add the line/column
location so it can be easily cross-references to the source code.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Fri, 28 Jun 2019 09:36:49 +0000 (19:36 +1000)]
parsergen: only non-terminals should make a state "starts_line"
If a state is followed by NEWLINE, then it isn't starts_line - more like
ends_line.
It is only non-terminals containing NEWLINEs that cause a state
to be starts_line.
So move the test to after we stop looking at terminals.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 23 Jun 2019 05:37:50 +0000 (15:37 +1000)]
oceani: allow 'then' in simple if statements.
Allow 'then' after "if expression", and don't require a ':' if
it is followed by simple statements.
Similarly "else" doesn't need a colon for simple statements
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 23 Jun 2019 04:41:47 +0000 (14:41 +1000)]
oceani: change parsing for ; at end
When we have 'for' and 'then' on the same line, I want to
require a ';' for the 'for' (and 'while').
So change SimpleStatemnts to never end with ';', and require
a ; or Newline after each instance of SimpleStatements.
Hmm... I guess that could be abstracted.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 23 Jun 2019 04:29:13 +0000 (14:29 +1000)]
oceani: modify grammar to not waste stack on newlines
Current grammar uses one stack frame per newline for leading
newlines as these productions are right-recursive. This is
unnecessary and inelegant. Change to use a left-recursive Newlines
production.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 23 Jun 2019 03:51:46 +0000 (13:51 +1000)]
indent_test: reduce stack usage for preceding NEWLINEs
In the cases where we allow preceding newlines (Statementlist Open Close)
we current use one parse-stack from for each newline. While there are
unlikely to be many, this is inelegant.
Change the right-recursive form to use a left-recursive Newlines rule
that absorbs one or more NEWLINEs using at most 2 stack frames.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 23 Jun 2019 00:21:14 +0000 (10:21 +1000)]
parsergen: allow $$OUT to be satisfied are start-of-line.
If a $$OUT (or $$NEWLINE) production is being reduced at
start-of-line (with no indents), then that is satisfactory,
we don't need NEWLINE etc as look-ahead.
This means that in cases where this is relevant, the computed
lookahead is wrong - we shouldn't have striped it.
I don't think this matters as it only affects conflict warnings,
and I think these will be reported at a higher level if relevant.
If essense, the $$OUT marking is like a precendence marking which
suppresses shift/reduce warnings as it say that decision is being made
on some basis other than look-ahead.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 22 Jun 2019 23:00:06 +0000 (09:00 +1000)]
parsergen: allow $$OUT as well as $$NEWLINE
Both of these mean the same thing: the product must end
in a 2-d visible way
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 22 Jun 2019 22:56:15 +0000 (08:56 +1000)]
parsergen: don't assume TK_newline in look-ahead for lineline
A linelike production might not have TK_newline in the look-ahead,
so don't assume it. If it isn't there, make the look-ahead
set empty.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 16 Jun 2019 02:37:23 +0000 (12:37 +1000)]
parsergen: revert the allowance for "non-critical" conflicts.
Having improved the parsing approach, I don't get any
"non-critical" conflicts any more, and I no longer think
they aren't critical.
Conflicts can often be resolved with precendence if they cannot
be resolved any other way.
So treat all conflict, except shift/reduce on NEWLINE, as genuine
conflicts.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 16 Jun 2019 02:34:40 +0000 (12:34 +1000)]
oceani-tests: add tests with lots of blank lines.
This found some problems with the grammer, which I fixed.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 16 Jun 2019 02:16:44 +0000 (12:16 +1000)]
oceani: convert declarations to new approach to newlines
There are now no conflicts. Yay!!
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 16 Jun 2019 01:37:03 +0000 (11:37 +1000)]
oceani: re-indent DeclareStruct
It wasn't indented to easily allow alternate productions.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 16 Jun 2019 01:35:00 +0000 (11:35 +1000)]
gitignore - add coverage_scanner
This is a tmp file - should have been ignored before.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 16 Jun 2019 01:31:54 +0000 (11:31 +1000)]
parsegen: fix up look-ahead for $$NEWLINE items.
I was discarding all non-newlines from the lookahead
in the wrong place.
I need to do it based on the productions added, not
the item the are generated by.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 16 Jun 2019 00:16:07 +0000 (10:16 +1000)]
oceani: change NEWLINE parsing in statements to new model.
The new module is:
A list of line-like things must access a newline first:
Statementlist -> Statements
| NEWLINE Statementlist
Any line-like thing must reduce to a single symbol:
SimpleStatementLine IfPart WhilePart CasePart etc
An individual line-like thing must allow following newlines
e.g.
IfHead -> if Expression Block
| IfHead NEWLINE
A block that can be multi-line or single-line should be marked with
$$NEWLINE
This will require a NEWLINE to reduce it, but won't swollow the newline.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 15 Jun 2019 23:47:58 +0000 (09:47 +1000)]
indent_test: adjust grammer to handle blank lines better.
This uses the new $$NEWLINE are other techniques to ensure
blank lines are handled well.
We also test that adding blank lines everywhere doesn't break
anything.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 15 Jun 2019 23:20:30 +0000 (09:20 +1000)]
ident_test: declare precedence for 'else'
By declaring precedence for 'else', we suppress conflict warnings
Normally newlines and indents will resolve any conflict, but
if not, else associates to the right - it should be shifted, not cause
a reduce (which is the default anyway)
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 15 Jun 2019 23:17:58 +0000 (09:17 +1000)]
indent_test: make testing easier - and include expected output
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 15 Jun 2019 23:14:03 +0000 (09:14 +1000)]
parsergen: allow non-virtual symbol to carry precedence.
When a production has $$foo to explicity set precedence, the 'foo'
doesn't really need to be virtual, it just needs to actually
carry a precedence.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 15 Jun 2019 22:29:16 +0000 (08:29 +1000)]
parsergen: introuce $$NEWINE pseudo-precedence.
Sometimes we need a produce to be terminated by a newline, but we
don't want to consume the newline with a "shift".
Case in point is:
Block -> : StatementList
Which can be used with
Statement -> if Expression Block
StatementList -> Statement
I want this to parse:
if something: if otherthing: action
which might seem a little odd, but is syntactically sensible.
The NEWLINE at the end is requred, and must close both nested Statements.
The NEWLINE will already cause a REDUCE, but if we don't have
Block -> : Statementlist NEWLINE
then something else could force a reduce, and we don't want that.
So introduce a marking "$$NEWLINE" which is similar to imposing a precedence
on a production. Now
Block -> : StatementList $$NEWLINE
means that a NEWLINE is required to end a Block, but it isn't
shifted. If anything else if found here, it is an error.
We also allow $eof and OUT to reduce this production.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sat, 15 Jun 2019 22:25:23 +0000 (08:25 +1000)]
parsergen: allow $<1 to be used for structs as well as pointer.
A future patch will want to use this for non-pointers, so make it
work - it is more consistent this way anyway.
Also: update to parsers to use $< where possible.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Mon, 10 Jun 2019 06:51:31 +0000 (16:51 +1000)]
Oceani: be sure to set precedence on required terminals.
In some cases I was only setting on nonterminals which derived the
terminals. That set it for the reduction, but not the terminal.
Also remove an ERROR production that duplicates one from Declaration
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Mon, 10 Jun 2019 06:34:55 +0000 (16:34 +1000)]
parsergen: don't report conflicts resolved by precedence.
If precedence information has been given which can resolve
a conflict, then don't report it.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 9 Jun 2019 23:11:05 +0000 (09:11 +1000)]
oceani: convert expression parsing to use precedences.
As precedences are now working (and largely have been for a long time).
it is time to start using them for oceani.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 9 Jun 2019 23:00:43 +0000 (09:00 +1000)]
parsergen: flip ordering of precedence declarations.
Change so first precedence declaration is the lowest precedence.
This is consistent with bison, and will make converting
'oceani' easer.
When not using precedence, it is easier to do the lowest
precedence first - so keep for approach.
Signed-off-by: NeilBrown <neil@brown.name>
NeilBrown [Sun, 9 Jun 2019 22:49:08 +0000 (08:49 +1000)]
oceani: use 'bracket' printing for expressions.
Adding brackets to expression printing removes and ambiguity.
As I'm about to change expression parsing, I want to be able to
see that the result is correct.
Signed-off-by: NeilBrown <neil@brown.name>