Re: [9fans] Abaco: url parsing problem - Plan9

This is a discussion on Re: [9fans] Abaco: url parsing problem - Plan9 ; Uriel wrote: >That is why somebody (cough) riped out the url parsing code from webfs >and put it into a library, and then updated abaco to use that >library... but we all know here code and effort duplication is good, ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: Re: [9fans] Abaco: url parsing problem

  1. Re: [9fans] Abaco: url parsing problem

    Uriel wrote:

    >That is why somebody (cough) riped out the url parsing code from webfs
    >and put it into a library, and then updated abaco to use that
    >library... but we all know here code and effort duplication is good,
    >so never mind.
    >
    >

    the browser needs to understand url-encoding anyway beacause he
    he needs to touch the url (build absolute urls from relatives in context
    of the fetched
    page). so having the lib in abaco is ok i think.

    but maybe other stuff could gain from webfs *fixing* urls. maybe some
    rc-scripts where you
    dont want to urlencode stuff first? for example you want to fetch a
    search-result page
    and dont want to urlencode your concatenated search query before passing
    it to webfs?

    makes that sense? ;-)

    >uriel
    >
    >On Dec 20, 2007 4:27 PM, Federico G. Benavento wrote:
    >
    >
    >>>Does webfs do the url parsing? or abaco before sending it to webfs, if so, maybe
    >>>the function can be fixed easily.
    >>>
    >>>

    >>I guess, my urlencode is broken, I'll fix it later
    >>
    >>thanks
    >>
    >>Federico G. Benavento
    >>
    >>---
    >>/bin/fortune:
    >>
    >>
    >>Even the simplest solution is bound to have something wrong with it.
    >>
    >>

    --
    cinap


  2. Re: [9fans] Abaco: url parsing problem

    On Dec 21, 2007 11:13 AM, Kernel Panic wrote:
    > Uriel wrote:
    >
    > >That is why somebody (cough) riped out the url parsing code from webfs
    > >and put it into a library, and then updated abaco to use that
    > >library... but we all know here code and effort duplication is good,
    > >so never mind.
    > >
    > >

    > the browser needs to understand url-encoding anyway beacause he
    > he needs to touch the url (build absolute urls from relatives in context
    > of the fetched
    > page). so having the lib in abaco is ok i think.


    Lets not forget javascript and other such abominations.

    And that is *precisely* why having two half broken sets of URL
    parsing/manipulation code, one in webfs and one in abaco is so silly.
    And why I factored out the url code in webfs and put it into a liburl,
    and then modified abaco link against that instead of its own
    simplistic and broken URL code.

    The result has been on my sources/contrib dir for ages, why fgb has
    not used it is beyond me.

    uriel


    > but maybe other stuff could gain from webfs *fixing* urls. maybe some
    > rc-scripts where you
    > dont want to urlencode stuff first? for example you want to fetch a
    > search-result page
    > and dont want to urlencode your concatenated search query before passing
    > it to webfs?
    >
    > makes that sense? ;-)
    >
    >
    > >uriel
    > >
    > >On Dec 20, 2007 4:27 PM, Federico G. Benavento wrote:
    > >
    > >
    > >>>Does webfs do the url parsing? or abaco before sending it to webfs, if so, maybe
    > >>>the function can be fixed easily.
    > >>>
    > >>>
    > >>I guess, my urlencode is broken, I'll fix it later
    > >>
    > >>thanks
    > >>
    > >>Federico G. Benavento
    > >>
    > >>---
    > >>/bin/fortune:
    > >>
    > >>
    > >>Even the simplest solution is bound to have something wrong with it.
    > >>
    > >>

    > --
    > cinap
    >
    >


  3. Re: [9fans] Abaco: url parsing problem

    On 21/12/2007, Uriel wrote:

    > The result has been on my sources/contrib dir for ages, why fgb has
    > not used it is beyond me.


    I will have a look at the library, I had to "partially" implement one
    for a web app
    that needed to save some data into a database from a cgi. It can be useful
    for me to see another implementation.

    --
    Fidonet: 2:345/432.2

+ Reply to Thread