Rmk193 fix some hcfiles font errors#2637
Conversation
Fixes the HCFILES fails for postscript fonts
Brute force fix to a few other HCFILES errors
|
On Linux Mint 22.1 Cinnamon I ran do_hcfiles.sh locally and, unlike files.interlisp.org, the loops and notecards directories are missing altogether. This is the log file: hcfiles-fails.txt |
|
At commit 7ee6c72 the directories loops and notecards are still missing. The fails file: hcfiles-fails.txt |
|
notecards and loops are separate git repositories that you can check out so their directories are siblings of the medley directory - then the hcfiles script should find them. |
|
In the failure file produced by the script, the notecards and loops files show up as subdirectories of {MEDLEY}. But they aren't in the same repo, and at least in my set up I have them as sister directories of {MEDLEY} (with pseudohosts {GLOOPS} and {GNOTECARDS} that GITFNS creates if it can find them).
… On Jun 4, 2026, at 9:37 AM, Paolo Amoroso ***@***.***> wrote:
pamoroso
left a comment
(Interlisp/medley#2637)
<#2637 (comment)>
At commit 7ee6c72 <7ee6c72> the directories loops and notecards are still missing. The fails file: hcfiles-fails.txt <https://github.com/user-attachments/files/28604458/hcfiles-fails.txt>
—
Reply to this email directly, view it on GitHub <#2637?email_source=notifications&email_token=AQSTUJIGPIGN2PJD3GUDSRT46GQWHA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINRSGQZDANBVGIY2M4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L#issuecomment-4624204521>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJL4PPQO5TE3O7ZFCCT46GQWHAVCNFSM6AAAAACZZ32QEKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DMMRUGIYDINJSGE>.
You are receiving this because you authored the thread.
|
|
The directories loops and notecards are actually already siblings of the medley directory, so I wonder why do_hcfiles.sh doesn't detect them: |
|
At commit a617708 I still don't get the directories loops and notecards. The fails file: hcfiles-fails.txt |
|
I assume you also don't get those in master
… On Jun 5, 2026, at 1:40 AM, Paolo Amoroso ***@***.***> wrote:
pamoroso
left a comment
(Interlisp/medley#2637)
<#2637 (comment)>
At commit a617708 <a617708> I still don't get the directories loops and notecards. The fails file: hcfiles-fails.txt <https://github.com/user-attachments/files/28630978/hcfiles-fails.txt>
—
Reply to this email directly, view it on GitHub <#2637?email_source=notifications&email_token=AQSTUJJDMMEJJ26CRBORTH346KBRFA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINRSHE3DSMZRGUY2M4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L#issuecomment-4629693151>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJPO7JBGEOK5YL6HTC346KBRFAVCNFSM6AAAAACZZ32QEKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DMMRZGY4TGMJVGE>.
You are receiving this because you authored the thread.
|
|
PR #2532 was an attempt to reduce some of the confusion around the sub-module / sister repository. Right now, we have things that are created by scripts in the medley repository that refer to sources not in the medley repository. in particular, the HCFILES script copies these "ordinarily sister" repositories to be under {medley}. I marked #2532 as draft with the idea of doing one sub-module repository at a time. |
|
@rmkaplan Yes, the loops and notecards directories are missing also in master. The fail file of master: hcfiles-fails.txt @masinter I don't know, so far I checked out submodules only lightly and find the feature confusing. |
|
I'm confused. It appears that the font 13 files in the bigger hcfiles-fails.txt are in the user-guide-v1.2> from_envos subdirectory, but there are versions of those files above in the user-guide-v1.2. And the versions higher up have already had the CR/LF swap (with a little fixing shell script that is stored at the docs/ level). So, why is HCFILES looking at the from_envos subdirectory? Maybe that directory should be moved to obsolete so we don't waste time on it. |
|
If GITFNS is loaded, it looks around for loops and notecards and tries to create pseudohosts {GNOTECARDS} and {GLOOPS} that point to them.
I don't know if its current search covers both the sister and subdirectory possibiliites. I have the sister configuration, and that works. The subdirectory needs to be tested.
If we verify that it covers both configurations, then HCFILES could use {GNOTECARDS} and {GLOOPS} and the paths listed in the HCFAILS file would be good in both environments.
… On Jun 5, 2026, at 8:39 AM, Larry Masinter ***@***.***> wrote:
masinter
left a comment
(Interlisp/medley#2637)
<#2637 (comment)>
PR #2532 <#2532> was an attempt to reduce some of the confusion around the sub-module / sister repository. Right now, we have things that are created by scripts in the medley repository that refer to sources not in the medley repository. in particular, the HCFILES script copies these "ordinarily sister" repositories to be under {medley}.
I marked #2532 <#2532> as draft with the idea of doing one sub-module repository at a time.
This will lead to an interim state where the module structure is half-done. Is that OK with you? @rmkaplan <https://github.com/rmkaplan> @pamoroso <https://github.com/pamoroso>
—
Reply to this email directly, view it on GitHub <#2637?email_source=notifications&email_token=AQSTUJJWY5EEMEMYBFXL3ML46LSU5A5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINRTGMYTMMRSGA22M4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-4633162205>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJO63LLB5DEQ42TUVF346LSU5AVCNFSM6AAAAACZZ32QEKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DMMZTGE3DEMRQGU>.
You are receiving this because you were mentioned.
|
|
Did you catch it after the second commit that removed the MEDLEY-INIT-VARS entries?
… On Jun 8, 2026, at 12:21 PM, Nick Briggs ***@***.***> wrote:
@nbriggs approved this pull request.
Looks good.
—
Reply to this email directly, view it on GitHub <#2637?email_source=notifications&email_token=AQSTUJPBUSXASKRKJRYHAJT464G3LA5CNFSNUABKM5UWIORPF5TWS5BNNB2WEL2QOVWGYUTFOF2WK43UKJSXM2LFO4XTINBVGI3TCMJXGY32M4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#pullrequestreview-4452711767>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJOA43GPALIGHUPSP5L464G3LAVCNFSM6AAAAACZZ32QEKVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHM2DINJSG4YTCNZWG4>.
You are receiving this because you were mentioned.
|
|
Yes. It would have shown as has having been changed since I approved it and would require reapproval if I'd jumped the gun on that. |
4 participants
The postscript failures in #2635 were the result of an argument mismatch between the POSTSCRIPTSTREAM call \SEARCHFONTFILES and the expectations of that function.
This PR also provides a brute-force patch for the DISPLAY failures for HELVETICA 6 and TIMESROMAN 24. Basicallly, asserting that TIMESROMAN can be coerced to CLASSIC, and that HELVETICA 6 can be coerced to HELVETICA 7.
Still don't know why it needs to create those display fonts when it is creating the image file, but if it asks for them, it will get a good approximation.
I tried coercing the size 13 fonts to 12, but that didn't help--it got an array error saying that an index 13 was out of bounds in a situation where 10 was perfectly fine. That reinforces the belief that there was a 10/13 swap in those notecards files--maybe byte swapping would fix those.