prevent oe from parsing documents?

Author Message
Mikk 09/11/2003 05:31 pm
is there a way to prevent oe from parsing .txt, .doc, .pdf, ... files?
i have a project which has >40.000 documents (txt, doc, pdf) which oe parses everytime i update the project although i know that there are no links in this files. this is of course slowing down the update of the project enormously.

it would be great if an option for this could be added.

perhaps it might even be better to make oe so that files are only parsed once and if no links are found the files are marked in the project-file so that it will not be parsed again unless it was updated on the server.
this way almost every project update could be speeded up greatly at the second (and higher) run.
Oleg Chernavin 09/12/2003 08:09 am
Thank you for the idea. I will see how it could be implemented (when the program cannot find links in a file). The only problem I see is that if some file was parsed and no links were found, but the next version adds support for that kind of file - what to do? Because even the new version will skip parsing these files.

Best regards,
Oleg Chernavin
MetaProducts corp.
Mikk 09/12/2003 08:44 am
> The only problem I see is that if some file was parsed and no links were found, but the next version adds support for that kind of file - what to do?

i?m not shure i understand what you mean with this. could you explain this a little bit.

thanks ,
Mikk
Oleg Chernavin 09/12/2003 09:50 am
OK. For example, there is a file.ext. OE 2.9 loads it, finds no link there and stores on the disk. The file.ext is marked as "Do not parse". Then I release 3.2 version after some time. 3.2 version can detect new kinds of links that 2.9 was unable to find in files. So, if 3.2 version would parse file.ext, it could find links there. But since the file was marked as "Do not parse" by a previous version, the new version will not even try to find links in it too.

Oleg.
Mikk 09/12/2003 11:50 am
> OK. For example, there is a file.ext. OE 2.9 loads it, finds no link there and stores on the disk. The file.ext is marked as "Do not parse". Then I release 3.2 version after some time. 3.2 version can detect new kinds of links that 2.9 was unable to find in files. So, if 3.2 version would parse file.ext, it could find links there. But since the file was marked as "Do not parse" by a previous version, the new version will not even try to find links in it too.
>
> Oleg.

ok thanks for explaining, now i understand.
what do you think about this in your example:
there is a file.ext.
OE 2.9 loads it, finds no link there and stores on the disk.
The file.ext is marked as "Do not parse, 2.9"

...-> new version 3.2 which can detect new kinds of links:

OE 3.2 first project update starts:
files are parsed -> OE 3.2 comes to file.ext (marked as "Do not parse, 2.9")
the mark 2.9 tells OE 3.2 that this file needs to be parsed again (3.2>2.9)
file is parsed and marked as "Do not parse, 3.2" (if no links are found of course)

OE 3.2 2nd project update starts:
files are parsed -> OE 3.2 comes to file.ext (marked as "Do not parse, 3.2")
the mark 3.2 tells OE 3.2 that this file needs NOT to be parsed again (3.2=3.2)
file is skipped (and a lot of time is saved in big projects)

with this files would be parsed once with any major update but would be left out after the first update with the same version.

even better would be perhaps if the files are not marked with the version number of the OE program but with an version of link detection.
for example:
OE 2.9
parses file and marks as "Do not parse, 1"

OE 3.0 (no change in link detection)
parses file and marks as "Do not parse, 1"

OE 3.1 (no change in link detection)
parses file and marks as "Do not parse, 1"

OE 3.2 (link detection updated)
parses file and marks as "Do not parse, 2"

as i?m no programmer i?m not completely shure about this but perhaps it will work this way (?)

regards,
Mikk
Mikk 09/12/2003 11:54 am
just found an mistake in my own post:

>OE 3.0 (no change in link detection)
>parses file and marks as "Do not parse, 1"
>
>OE 3.1 (no change in link detection)
>parses file and marks as "Do not parse, 1"

these two versions would of course not parse the file as the link detection wasn?t changed.
Oleg Chernavin 09/15/2003 04:58 am
Yes, I see the logic here. I will think more about it.

Oleg.