Enable CTL for Default?

[Joel Madero]

If we can run a test to look at performance and verify that it's
minimal or no inpact maybe we can turn CTL on by default but leave the
option to turn it off. Let our team play around with it for awhile
before submitting it to the master branch. This way if it does cause a
major headache at least:

a) most users won't see it
b) we can report/diagnose "in house"
c) We know that it's easily fixable by just disabling CTL


Just a thought.


Joel

On Tue, Jun 19, 2012 at 11:24 AM, Norbert Thiebaud <nthiebaud at gmail.com> wrote:
> On Tue, Jun 19, 2012 at 1:12 PM, Christian Lohmaier
> <lohmaier+libreoffice at googlemail.com> wrote:
>> Hi Norbert,  *,
>>
>> On Tue, Jun 19, 2012 at 7:52 PM, Norbert Thiebaud <nthiebaud at gmail.com> wrote:
>>> On Tue, Jun 19, 2012 at 8:41 AM, Christian Lohmaier
>>> <lohmaier+libreoffice at googlemail.com> wrote:
>>> [..]
>>>> * visible cursor travelling, input sequence checking and detection of
>>>> writing mode that are only active for CTL might have a tiny
>>>> performance impact
>>>
>>> Either it _has_ a tiny performance impact
>>> or it _might_ have a performance impact
>>> But qualifying a performance impact, that has not been measured, as
>>> 'tiny' up-front, is 'begging the question'.
>>
>> Well, I have a fairly slow machine, and I have enable Asian language
>> support enabled and don't notice any difference compared to not having
>> it enabled. But I don't work with huge documents.
>> Also I frequently did enable CTL support when I was doing more QA work
>> (processing issues) and didn't note any performance drawbacks either,
>> but then again I only used it to check/verify bugs, not to actually
>> create documents that are RTL or whatnot.
>>
>> I only use the textgrid and vertical writing myself.
>>
>>> (note: it is quite possible that the impact, if any, is
>>> tiny/negligible... I'm just objecting to the biased wording of the
>>> statement of 'fact')
>>
>> Well, free free to provide some numbers, or even a way to measure it.
>> If stuff is done in addition, that needs more CPU. That is my
>> rationale. I myself didn't notice any penalty that would make me not
>> activate the option because of performance issues, but I'm no heavy
>> user myself.
>>
>
> Fair enough.
>
>>>> * will work out-of the box for all users
>>> No regression for any document that does not rely on these features today ?
>>> (note: This is not a rhetorical question. I honestly do not know if
>>> these options only impact the UI or if they also can impact the way a
>>> document is rendered)
>>
>> Well - I always was under the assumption that the option would only
>> affect the UI, but Eike corrected me on that point and mentioned the
>> additional input sequence checking and writing mode detection & cursor
>> traveling stuff that is only used with the option enabled.
>>
>> If there is a change in rendering of the document based on whether the
>> option is enabled or not, then it is a bug.
>
> I agree it would be... I'm concerned that they would all become very
> visible at once...
>
> Norbert