I have a script that finds and reads and processes a "saved as" Firefox file then deletes it and wait for a new "saved as" file of the same name but although the old file has been effectively "deleted" (its name is not there anymore in "ls" before getting the new file), the script then processes the contents of the old file, not the contents of the new file and no matter of how many times this new file is changed. I suspect it is fetching the old original file contents of the same old file from a cache somewhere (since the name is not changed).
Could someone make suggestions (clear cache or use command more appropriate than "rm" or another browser)?

Using Firefox-esr 60.2.2

Thank you for your help.

a man walks in to the doctors office, and says, hey Doc. I got a freind that every time he stands up, his pants falls down. What do you think is the matter with him?

Doc: Maybe he needs a new belt.
Friend: but he has no hips to keep the belt up.

gee I wish I had something to look at so I could tell what is really going on.

1 members found this post helpful.

If you edit a file and want to save the revised version by replacing the original, you should use the SAVE button. The original file will be replaced by the revised version.

If you want to save the revised version with a different file name, you should use the SAVE AS button, and (when prompted) give the name for the new file. The original file will remain as it was.

Your post prompts the thought that you are using the wrong button and Firefox is operating correctly.

Daniel B. Martin

.

1 members found this post helpful.

The file is saved "as is" and then the "saved" file (its copy) is read by script and processed, not the original but Firefox does not offer a "save" option in this case, only a "save as" option. After processing, the "saved as" file (not the original) is removed (rm) and Firefox is closed. Finding the file is done by using "find newer.." than a dummy file just "touched" by the script for that purpose. After a certain time, the problem automatically disappears - for example waiting long enough between the processing of 2 different files by the same instance of the script which seems to support my view that a cache is involved.

It would be easier to troubleshoot this if you posted your actual script, or the relevant section(s) thereof.

1 members found this post helpful.

It's not enough to delete (unlink) the file, you have to close (close/fclose) it, then you can call open/fopen again.

2 members found this post helpful.

Yes. It's not what Firefox does, but what the script does with the file.

1 members found this post helpful.

Most likely unrelated, but once I had a mishap with glibc that did cause problems in firefox, see here:
https://bugzilla.mozilla.org/show_bu...=1042134#cs-26

(Firefox dlopened 'libc.so.6' and ended up with two libc.so.6 dynamically loaded... that had disturbing results.)

1 members found this post helpful.

Checked and files are properly open and closed at the appropriate time (just before and after reading/writing), I was expecting an obvious trick, not having much experience dealing with browser files, it is not the case. The script is over 900 lines of code, posting it will be too much for the volunteers to investigate and posting the relevant code will always leave the doubt there is something else of importance not showing. I am also limited with time and found the easy solution is to give the files a different name/extension (.001 etc) which works.