Add Rez

[cooljeanius]

Description

See issue #6107

Checklist:

• I am adding a new language.
  • The extension of the new language is used in hundreds of repositories on GitHub.com.
    • Search results for each extension:
      • https://github.com/search?q=path%3A*.r+DITL&type=code
      • https://github.com/search?q=path%3A*.r+verUS&type=code
      • https://github.com/search?q=path%3A*.r+purgeable&type=code
  • I have included a real-world usage sample for all extensions added in this PR:
    • Sample source(s):
      • https://github.com/cy384/ssheven/blob/master/ssheven-icons.r
      • https://github.com/cy384/ssheven/blob/master/ssheven.r
      • https://github.com/chrisballinger/rezilla/blob/72c4cfd1a2d45565d2b429bd6afff8bf426f5c65/RezillaSource/Rezilla_Rsrc/Rezilla_DITL.r
      • https://github.com/autc04/Retro68/blob/5f882506013a0a8a4335350197a1b7c91763494e/Samples/Dialog/dialog.r
      • https://github.com/vim/vim-history/blob/868c2742154cbb6d59bfd2818724b980c50eb24b/src/gui_mac.r
    • Sample license(s):
      • https://github.com/cy384/ssheven/blob/master/LICENSE
      • https://github.com/autc04/Retro68/blob/master/COPYING
      • https://github.com/vim/vim/blob/master/LICENSE
  • I have included a syntax highlighting grammar: https://github.com/textmate/rez.tmbundle
  • I have added a color
    • Hex value: #FFDAB3
    • Rationale: It's the most commonly-used color in the ResEdit icon that isn't confusable with R (see comment from @uliwitness below)
  • I have updated the heuristics to distinguish my language from others using the same extension. (WIP)

[uliwitness]

If you're looking for a color: R has blue and grey, so I'd say pick one of the remaining colors from the ResEdit icon, which seem to be green (00CE00), red (E60100), or a light salmon ("flesh") color (FFDAB3).

For detecting, the most common resource types are probably 'vers', 'SIZE', 'ICON', 'cicn', 'ICN#', 'icns', 'BNDL', 'DITL', 'DLOG', 'ALRT', 'MENU', 'CURS' and 'PICT'. Might also be worth looking for resource '****' (with **** being a pattern, 4 characters) and data '****', and maybe #include <Types.r> and #include <Carbon/Carbon.r>.

[cooljeanius]

If you're looking for a color: R has blue and grey, so I'd say pick one of the remaining colors from the ResEdit icon, which seem to be green (00CE00), red (E60100), or a light salmon ("flesh") color (FFDAB3).

OK, I counted how many pixels of each there were in the icon:

• green (both 00CE00 and a slightly different shade): 7
• E60100: 24
• FFDAB3: 44

This suggests to me that we should go with the last of the 3.

For detecting, the most common resource types are probably 'vers', 'SIZE', 'ICON', 'cicn', 'ICN#', 'icns', 'BNDL', 'DITL', 'DLOG', 'ALRT', 'MENU', 'CURS' and 'PICT'. Might also be worth looking for resource '****' (with **** being a pattern, 4 characters) and data '****', and maybe #include <Types.r> and #include <Carbon/Carbon.r>.

OK, could I get some help with a regex to catch all of those? I'm not very good with regexes...

[uliwitness]

Actually, this might be a better heuristic:

(resource|data|type)\s+'[A-Za-z0-9]{4}'\s+((\(.*\)\s){0,1})+{

That will basically catch the first line of a resource declaration, as long as they don't have extra line breaks.

[uliwitness]

For detecting some common types, one would do

'(ICON|BNDL|vers|SIZE|cicn|ICN#|icns|DITL|DLOG|ALRT|MENU|MBAR|CURS|PICT|cfrg|thng)'

[uliwitness]

To detect the includes one would do

#include\s+["<](Types\.r|Carbon/Carbon\.r)[">]

I guess detecting any of these three regexes could be seen to indicate that it's a Rez file. So maybe something like:

(#include\s+["<](Types\.r|Carbon/Carbon\.r)[">])|('(ICON|BNDL|vers|SIZE|cicn|ICN#|icns|DITL|DLOG|ALRT|MENU|MBAR|CURS|PICT|cfrg|thng)')|((resource|data|type)\s+'[A-Za-z0-9]{4}'\s+((\(.*\)\s){0,1})+{)

should probably catch something in most files. Though some of these might give false positives if the files don't have the '.r' suffix. That said, I haven't seen an R source file yet, so we might wanna check that none of these are valid/common in that other language.

[cooljeanius]

To detect the includes one would do

#include\s+["<](Types\.r|Carbon/Carbon\.r)[">]

I guess detecting any of these three regexes could be seen to indicate that it's a Rez file. So maybe something like:

(#include\s+["<](Types\.r|Carbon/Carbon\.r)[">])|('(ICON|BNDL|vers|SIZE|cicn|ICN#|icns|DITL|DLOG|ALRT|MENU|MBAR|CURS|PICT|cfrg|thng)')|((resource|data|type)\s+'[A-Za-z0-9]{4}'\s+((\(.*\)\s){0,1})+{)

should probably catch something in most files. Though some of these might give false positives if the files don't have the '.r' suffix. That said, I haven't seen an R source file yet, so we might wanna check that none of these are valid/common in that other language.

There are some examples of R sources in linguist's samples directory, if you want to take a look

[uliwitness]

OK, the import.r sample file in there has single-quoted strings like 'name'. While I haven't seen 'ICON' or the like yet, we should probably remove that part from the regex and just look for actual syntax, conservatively:

(#include\s+["<](Types\.r|Carbon/Carbon\.r)[">])|((resource|data|type)\s+'[A-Za-z0-9]{4}'\s+((\(.*\)\s+){0,1}){)

[uliwitness]

Oh, I also think I overlooked how one specifies a resource name in Rez.

Update: It's fine, the name is inside the parentheses where the ID is, so we correctly parse that case too: resource 'SIZE' (-1, "Our SIZE -1") {

[cooljeanius]

OK, the import.r sample file in there has single-quoted strings like 'name'. While I haven't seen 'ICON' or the like yet, we should probably remove that part from the regex and just look for actual syntax, conservatively:

(#include\s+["<](Types\.r|Carbon/Carbon\.r)[">])|((resource|data|type)\s+'[A-Za-z0-9]{4}'\s+((\(.*\)\s){0,1})+{)

ok, so does that need to be quoted, or can it just be pasted as-is? Generally the other regexes in heuristics.yml are surrounded by single-quotes, and I see single-quotes in that string, so I wonder if they'll need to be escaped?

[uliwitness]

Oh yeah since that uses YAML and that single-quotes these regexes, you'll prolly have to escape the backslashes and the single quotes, good catch. So the entire, quoted regex should be something like:

'(#include\\s+["<](Types\\.r|Carbon/Carbon\\.r)[">])|((resource|data|type)\\s+\'[A-Za-z0-9]{4}\'\\s+((\\(.*\\)\\s+){0,1}){)'

[cooljeanius]

Oh yeah since that uses YAML and that single-quotes these regexes, you'll prolly have to escape the backslashes and the single quotes, good catch. So the entire, quoted regex should be something like:

'(#include\\s+["<](Types\\.r|Carbon/Carbon\\.r)[">])|((resource|data|type)\\s+\'[A-Za-z0-9]{4}\'\\s+((\\(.*\\)\\s+){0,1}){)'

OK thanks, I made this the regex in 2837cbc. Unfortunately bundler is giving me some issues that are preventing me from testing locally, so hopefully a maintainer will let the CI workflows run...

[lildude]

See comments inline.

Please remove or replace the two large samples that are currently collapsed in the PR view; they're too big with the icons file mostly filled stuff that isn't relevant to the language.

We also need the cached license file the script/add-grammar should have downloaded when your added the grammar.

[cooljeanius]

See comments inline.

OK, addressed.

Please remove or replace the two large samples that are currently collapsed in the PR view; they're too big with the icons file mostly filled stuff that isn't relevant to the language.

Removed in 769cf15.

We also need the cached license file the script/add-grammar should have downloaded when your added the grammar.

Hm, I guess I should just re-run the script then, if I'm not seeing it?

[cooljeanius]

We also need the cached license file the script/add-grammar should have downloaded when your added the grammar.

Hm, I guess I should just re-run the script then, if I'm not seeing it?

OK, re-ran; is the file added in dadde18 what you were requesting?

[lildude]

OK, re-ran; is the file added in dadde18 what you were requesting?

Yup. That's the one.

[lildude]

I've rectified the last of the issues and all looks good now. Usage appears high enough for inclusion too 👍