Work on symbolic accesses to structure fields #162
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Researching how to do this will be generally useful for cleaning the codebase, and will help me be more confident in how to futz with this routine for semantic pagebreaking.
…eferences I've started exploring Coccinelle since it seems like it should be a great refactoring tool. But, I have a PhD in astrophysics and I still find the documentation to be confusing. I did get this example patch to work, though: ``` @@ expression E; @@ - mem[E + 1].b32.s1 = INF_PENALTY + PENALTY_NODE_penalty(E) = INF_PENALTY ``` Applied with: ``` spatch --sp-file structs.sp --in-place --dir tectonic ```
When Coccinelle works, it works well! ``` @@ expression E; identifier I =~ ".*_NODE"; @@ - mem[E].b16.s1 = I + NODE_type(E) = I @@ expression E; binary operator O; identifier I =~ ".*_NODE"; @@ - mem[E].b16.s1 O I + NODE_type(E) O I ``` Run as: ``` spatch --sp-file structs.sp --in-place --dir tectonic ```
**The following Coccinelle patch was over-eager** and matched some expressions that were clearly NODE_subtypes: ``` @@ expression E; binary operator O; @@ - mem[E].b16.s0 O NORMAL + GLUE_SPEC_shrink_order(E) O NORMAL @@ expression E; @@ - mem[E].b16.s0 = NORMAL + GLUE_SPEC_shrink_order(E) = NORMAL ```
Almost all usages of MIN_HALFWORD in the original WEB code come through the NULL macro, which defines a NULL "pointer" into the memory array. I've tried to preserve the places that genuinely use the MIN_HALFWORD symbolic constant; there are only a dozen or so.
Thanks Coccinelle! ``` @@ expression E; @@ - E >= hi_mem_min + is_char_node(E) ```
``` @@ expression E; @@ - f = mem[E + 1].b16.s1; + f = LIGATURE_NODE_lig_font(E); ```
``` @@ expression E; @@ - f = mem[E].b16.s1; + f = CHAR_NODE_font(E); ```
``` @@ expression E; @@ - c = mem[E].b16.s0; + c = CHAR_NODE_character(E); ```
``` @@ expression E1, E2, E3; @@ - effective_char(E1, E2, mem[E3 + 1].b16.s0) + effective_char(E1, E2, LIGATURE_NODE_lig_char(E3)) ```
``` @@ expression E1, E2, E3; @@ - effective_char(E1, E2, mem[E3].b16.s0) + effective_char(E1, E2, CHAR_NODE_character(E3)) ```
It's packed in a funky way. This is a start: ``` @@ expression E1, E2; @@ - font_info[char_base[E1] + E2] + FONT_CHARACTER_INFO(E1, E2) ```
We interact with the existing code more smoothly if the charinfo is the `memory_word.b32` member, not `memory_word` itself.
Some manual tweaking, plus: ``` @@ expression F, I; @@ - font_info[height_base[F] + (I.s2) / 16].b32.s1 + FONT_CHARINFO_HEIGHT(F, I) ```
Again: ``` @@ expression F, I; @@ - font_info[depth_base[F] + (I.s2) % 16].b32.s1 + FONT_CHARINFO_DEPTH(F, I) ```
Again: ``` @@ expression F, I; @@ - font_info[italic_base[F] + (I.s1) / 4].b32.s1 + FONT_CHARINFO_ITALCORR(F, I) ```
``` @@ expression E; @@ - delete_glue_ref(mem[E + 4].b32.s1) + delete_glue_ref(INSERTION_NODE_split_top_ptr(E)) ```
``` @@ expression E; @@ - mem[E + 6].gr + BOX_glue_set(E) ```
``` @@ expression E; binary operator O; @@ - mem[E].b16.s0 O EXPLICIT + NODE_subtype(E) O EXPLICIT @@ expression E; @@ - mem[E].b16.s0 = EXPLICIT; + NODE_subtype(E) = EXPLICIT; ```
With: ``` @@ expression E; binary operator O; @@ - mem[E].b16.s0 O SPACE_ADJUSTMENT + NODE_subtype(E) O SPACE_ADJUSTMENT @@ expression E; @@ - mem[E].b16.s0 = SPACE_ADJUSTMENT; + NODE_subtype(E) = SPACE_ADJUSTMENT; @@ expression E; binary operator O; @@ - mem[E].b16.s0 O ACC_KERN + NODE_subtype(E) O ACC_KERN @@ expression E; @@ - mem[E].b16.s0 = ACC_KERN; + NODE_subtype(E) = ACC_KERN; ```
|
Nice! 👍 |
Mrmaxmeier
pushed a commit
to Mrmaxmeier/tectonic
that referenced
this pull request
Oct 28, 2019
…or-486 xetex get_date_and_time: pull from SOURCE_DATE_EPOCH
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
To understand the
line_breakroutine so that I can futz with it, it helps a lot to turn opaquememory_wordreferences into named structure field accesss. As far as I can tell, this has to largely be done manually. So, let's just get started.Coccinelle helps a lot. Almost none of the field accesses are uniquely interpretable in the current code on their own, but contextual hints can be very useful for figuring out what's being done.